In [1]:
import os
import anndata

import numpy as np
import pandas as pd
import scanpy as sc
import seaborn as sns
import matplotlib.pyplot as plt
from matplotlib.colors import LinearSegmentedColormap

sc.settings.verbosity = 2           # verbosity: errors (0), warnings (1), info (2), hints (3)
sc.logging.print_header()
sc.settings.set_figure_params(dpi=90, facecolor='white')
np.random.seed(0)

scanpy==1.9.2 anndata==0.8.0 umap==0.5.3 numpy==1.21.6 scipy==1.9.1 pandas==1.5.3 scikit-learn==1.0.2 statsmodels==0.13.2 python-igraph==0.10.4 pynndescent==0.5.8
In [2]:
### Seurat style colormap
colors = ["magenta", "k", "yellow"]
seurat_style = LinearSegmentedColormap.from_list("seurat", colors)
In [3]:
plot_color = ['#B68136', '#A5FCA1', '#14B0A7', '#E1E79F', '#E78A5D', '#A64FB9', '#997CA4', '#4D14C3', '#F03132', '#DA3E89', '#9865E8', '#54ABBD', '#D1B2C1', '#B4788E', '#9BD87C', '#73A2ED', '#D1AEB6', '#52EE6F', '#0C8CBF', '#7DA2E2']
In [4]:
fig5path = './output_files_testXR/'
if not os.path.exists(fig5path):
    os.makedirs(fig5path)
spatial_aspect = 7/11

Loading data¶

In [5]:
# cbm_file = "./input_files/cellbits_matrix_mean_inten.zip"
cbm_file = "./input_files/Fig5_cell_bit_intensity_mat.csv"
meta_file = "./input_files/Fig5_cells_positions.csv"
cells_meta = pd.read_csv(meta_file, index_col = 0)
cbm = pd.read_csv(cbm_file, index_col=0)

display(cbm)
display(cells_meta)
display(cells_meta['QC'].value_counts())
M40 M37 M15 M47 M53 M21 M09 M34 M22 M49 ... M35 M31 M30 M24 M38 M39 M10 M26 M16 M08
0 55.294118 117.658824 49.270588 91.211765 95.435294 1523.835294 188.200000 685.176471 3490.058824 265.447059 ... 582.188235 563.258824 2566.305882 1818.552941 104.564706 174.670588 65.011765 571.494118 36.788235 111.541176
1 10.584906 25.396226 8.716981 23.830189 11.622642 39.660377 3.226415 60.603774 268.226415 14.452830 ... 532.962264 43.018868 76.132075 295.264151 18.094340 15.773585 4.226415 915.811321 4.811321 2.075472
2 27.150000 1082.216667 15.116667 7.383333 5.650000 115.933333 5.133333 31.583333 66.850000 5.816667 ... 32.100000 34.150000 157.500000 77.066667 14.683333 10.916667 2.850000 62.950000 4.750000 0.716667
3 34.267241 1701.353448 17.517241 16.155172 29.655172 162.198276 5.413793 32.224138 60.172414 14.568966 ... 43.000000 34.870690 146.120690 113.034483 14.362069 23.784483 4.008621 85.844828 3.896552 2.603448
4 67.293333 850.493333 5.960000 255.440000 181.573333 374.986667 50.546667 118.733333 635.306667 26.946667 ... 165.426667 70.640000 302.106667 115.880000 23.786667 82.613333 5.826667 126.080000 3.493333 4.880000
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
56347 27.040230 1849.408046 17.626437 84.913793 146.793103 255.402299 74.609195 100.678161 308.770115 58.109195 ... 313.522989 351.522989 2584.787356 464.683908 68.350575 87.747126 16.959770 458.114943 14.505747 13.873563
56348 49.674074 968.051852 11.555556 78.644444 92.311111 111.466667 33.822222 143.962963 336.977778 21.259259 ... 43.525926 28.362963 167.451852 141.177778 40.533333 79.614815 14.570370 88.466667 11.600000 6.925926
56349 26.243590 254.564103 10.576923 29.833333 73.692308 87.102564 32.756410 171.500000 200.038462 32.089744 ... 34.333333 41.564103 199.910256 539.564103 60.551282 83.615385 12.884615 58.487179 8.384615 6.487179
56350 35.562500 494.468750 11.531250 311.906250 297.500000 143.031250 131.093750 378.781250 875.781250 24.468750 ... 38.187500 34.250000 201.968750 190.062500 22.562500 85.625000 13.656250 127.125000 8.093750 3.968750
56351 13.837838 46.770270 11.594595 20.189189 33.310811 36.797297 13.229730 66.810811 85.905405 16.270270 ... 191.770270 43.567568 83.027027 147.081081 8.648649 22.270270 7.702703 142.648649 12.243243 7.662162

56096 rows × 53 columns

x y leiden QC
0 8670.835294 12770.305880 13 cell
1 7887.358491 13559.679250 7 cell
2 6411.150000 12702.433330 6 cell
3 6473.422414 12705.620690 3 cell
4 6685.226667 12703.000000 28 cell
... ... ... ... ...
56347 6250.522989 1264.034483 9 cell
56348 5606.659259 1264.325926 20 cell
56349 5464.730769 1266.589744 22 cell
56350 5897.812500 1268.093750 20 cell
56351 2386.581081 354.770270 21 cell

56239 rows × 4 columns

cell    56096
lowq      143
Name: QC, dtype: int64
In [6]:
cells_meta = cells_meta[cells_meta['QC'] != 'lowq'] 
cbm = cbm.loc[cells_meta.index.to_list(),:]

print(cells_meta.shape, cbm.shape)

(56096, 4) (56096, 53)

Filter cells with high expression across >30 modules¶

In [7]:
cutoff = np.percentile(cbm, 98, axis=0)
highexpr_in_n_modules = (cbm>cutoff).sum(axis=1)
suspicious_cells = cells_meta.loc[highexpr_in_n_modules>10, :]
display(suspicious_cells)
# plt.hist(highexpr_in_n_modules, bins=20)

plt.figure(figsize=(8,8))
plt.scatter(cells_meta.loc[:,"x"], cells_meta.loc[:, "y"], s = 0.1, c ='gray')
plt.scatter(suspicious_cells.loc[:,"x"], suspicious_cells.loc[:, "y"], s = 0.5, c = 'red')
plt.gca().set_aspect("equal")
plt.gca().invert_yaxis()
x y leiden QC
0 8670.835294 12770.305880 13 cell
932 5172.469697 13358.469700 4 cell
1076 6074.665188 13567.004430 24 cell
1442 4929.518987 12883.894510 4 cell
1539 4639.836957 13107.275360 7 cell
... ... ... ... ...
56287 5987.407534 1111.945205 2 cell
56288 6090.383333 1115.772222 2 cell
56299 6111.675214 1131.034188 2 cell
56303 6126.366548 1143.459075 2 cell
56342 6025.068376 1257.021368 2 cell

791 rows × 4 columns

In [8]:
# Do the filtering
cbm = cbm[~cbm.index.isin(suspicious_cells.index)]
cells_meta = cells_meta[~cells_meta.index.isin(suspicious_cells.index)]
print(cells_meta.shape, cbm.shape)

(55305, 4) (55305, 53)
In [9]:
plt.figure(figsize=(8,8))
plt.scatter(cells_meta.loc[:,"x"], cells_meta.loc[:, "y"], s = 0.1)
plt.gca().set_aspect("equal")
plt.gca().invert_yaxis()
In [10]:
### Filter cells with high expression across >30 modules
cutoff = np.percentile(cbm, 98, axis=0)
highexpr_in_n_modules = (cbm>cutoff).sum(axis=1)
suspicious_cells = cells_meta.loc[highexpr_in_n_modules>10, :]
display(suspicious_cells)

# Visualize suspicious cells in red
plt.figure(figsize=(8,8))
plt.scatter(cells_meta.loc[:,"x"], cells_meta.loc[:, "y"], s = 0.1, c ='gray')
plt.scatter(suspicious_cells.loc[:,"x"], suspicious_cells.loc[:, "y"], s = 0.5, c = 'red')
plt.gca().set_aspect("equal")
plt.gca().invert_yaxis()

# Do filtering
cbm = cbm[~cbm.index.isin(suspicious_cells.index)]
cells_meta = cells_meta[~cells_meta.index.isin(suspicious_cells.index)]
print(cells_meta.shape, cbm.shape)
x y leiden QC
926 5614.473684 13352.047850 4 cell
1428 4826.044118 12861.676470 4 cell
1450 4853.080000 12900.074290 4 cell
1594 4713.255556 13240.611110 4 cell
1602 4711.144330 13251.000000 4 cell
... ... ... ... ...
55741 8396.232653 2275.575510 2 cell
55783 7937.610000 2338.400000 2 cell
56054 6581.252830 928.562264 2 cell
56252 6319.231132 860.400943 14 cell
56260 6247.503226 932.529032 2 cell

471 rows × 4 columns

(54834, 4) (54834, 53)
In [ ]:

Clustering¶

In [11]:
adata = sc.AnnData(cbm, dtype=np.float32)
adata.obs = cells_meta
adata.var_names_make_unique() 
print(adata.obs)

                 x             y  leiden    QC
1      7887.358491  13559.679250       7  cell
2      6411.150000  12702.433330       6  cell
3      6473.422414  12705.620690       3  cell
4      6685.226667  12703.000000      28  cell
5      6696.637500  12704.762500       4  cell
...            ...           ...     ...   ...
56347  6250.522989   1264.034483       9  cell
56348  5606.659259   1264.325926      20  cell
56349  5464.730769   1266.589744      22  cell
56350  5897.812500   1268.093750      20  cell
56351  2386.581081    354.770270      21  cell

[54834 rows x 4 columns]

C:\Users\Xinrui\anaconda3\lib\site-packages\anndata\_core\anndata.py:121: ImplicitModificationWarning: Transforming to str index.
  warnings.warn("Transforming to str index.", ImplicitModificationWarning)
C:\Users\Xinrui\anaconda3\lib\site-packages\anndata\_core\anndata.py:798: UserWarning: 
AnnData expects .obs.index to contain strings, but got values like:
    [1, 2, 3, 4, 5]

    Inferred to be: integer

  value_idx = self._prep_dim_index(value.index, attr)
In [12]:
sc.pl.highest_expr_genes(adata, n_top=cbm.shape[1], )
sc.pp.filter_cells(adata, min_genes=0)
sc.pp.filter_genes(adata, min_cells=0)

normalizing counts per cell
    finished (0:00:00)

C:\Users\Xinrui\anaconda3\lib\site-packages\seaborn\categorical.py:82: FutureWarning: iteritems is deprecated and will be removed in a future version. Use .items instead.
  plot_data = [np.asarray(s, float) for k, s in iter_data]

C:\Users\Xinrui\anaconda3\lib\site-packages\anndata\_core\anndata.py:121: ImplicitModificationWarning: Transforming to str index.
  warnings.warn("Transforming to str index.", ImplicitModificationWarning)
In [13]:
sc.pp.normalize_total(adata, target_sum=1e4, key_added = 'normalize_total')
sc.pp.log1p(adata, copy = False)
sc.pp.scale(adata, zero_center = True, max_value=10, copy=False, layer=None, )

normalizing counts per cell
    finished (0:00:00)

Clustering¶

In [14]:
clusters_idv_ids_fn = os.path.join(fig5path, 'Clustermap_idv_ids.pdf')
clusters_idv_names_fn = os.path.join(fig5path, 'Clustermap_idv_names.pdf')
clusters_fused_fn = os.path.join(fig5path, 'Clustermap_fused.pdf')
clusters_fused_names_fn = os.path.join(fig5path, 'Clustermap_fused_names.pdf')
adata_fn = os.path.join(fig5path, 'Fig5_mouseBrain_10xFISHnCHIPs.h5ad')
umap_ids_fn = os.path.join(fig5path, 'UMAP_ids.pdf')
heatmap_ids_fn = os.path.join(fig5path, 'Heatmap_ids.pdf')
In [15]:
n_neighbors = 15
n_pcs = 20
res = 0.6

sc.tl.pca(adata, svd_solver='arpack', use_highly_variable=False, n_comps=50, zero_center =True)
sc.pp.neighbors(adata, n_neighbors=n_neighbors, n_pcs=n_pcs, knn=True, method='umap', metric='euclidean')
sc.tl.leiden(adata, resolution=res)

computing PCA
    with n_comps=50
    finished (0:00:00)
computing neighbors
    using 'X_pca' with n_pcs = 20
    finished (0:00:21)
running Leiden clustering
    finished (0:00:08)
In [16]:
# Plot clusters
cluster_name = 'leiden'

adata.obs[cluster_name] = adata.obs[cluster_name].astype('category')
cluster_order = np.unique(adata.obs.loc[:,cluster_name]) 
n_cluster=len(cluster_order)
color_map_cluster = {cluster_order[i]: plot_color[i] for i in range(len(cluster_order))}
print('%i clusters found!\n' % n_cluster)
print(color_map_cluster)

sns.set(rc={'axes.facecolor':'black', 'axes.grid': False, 'axes.axisbelow': False, 
            "xtick.labelbottom": False, 'ytick.labelleft': False, 
            'axes.linewidth': 1.0, 'axes.edgecolor': '.8'})
facet = sns.relplot(data=adata.obs, x="x", y="y",  palette=color_map_cluster,  col=cluster_name, hue=cluster_name,  
                    linewidth=0, col_wrap = 6, kind="scatter", s=3, aspect=spatial_aspect)
facet.fig.axes[0].invert_yaxis()
# plt.savefig(clusters_idv_ids_fn, bbox_inches='tight')


plt.figure(figsize=(12, 12))
sns.scatterplot(data = adata.obs, x = "x" , y = "y", s = 3, linewidth=0, hue= cluster_name, palette=color_map_cluster)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0, ncol=1)
plt.gca().set_aspect("equal")
plt.gca().invert_yaxis()
# plt.savefig(clusters_idv_names_fn, bbox_inches='tight')

18 clusters found!

{'0': '#B68136', '1': '#A5FCA1', '10': '#14B0A7', '11': '#E1E79F', '12': '#E78A5D', '13': '#A64FB9', '14': '#997CA4', '15': '#4D14C3', '16': '#F03132', '17': '#DA3E89', '2': '#9865E8', '3': '#54ABBD', '4': '#D1B2C1', '5': '#B4788E', '6': '#9BD87C', '7': '#73A2ED', '8': '#D1AEB6', '9': '#52EE6F'}
In [17]:
# Plot UMAP
min_dist = 0.1
layer = 'leiden'
sc.tl.umap(adata, min_dist=min_dist, spread=1, n_components=2, maxiter=None, init_pos='spectral', random_state=0, copy=False)
sns.set(rc={'axes.facecolor':'white', 'axes.grid': False, 'figure.figsize': (10,10)})
# sc.pl.umap(adata, color=layer, s=8, palette=color_map_cluster, legend_loc='on data', save='Fig5_UMAP_id.pdf')
sc.pl.umap(adata, color=layer, s=8, palette=color_map_cluster, legend_loc='on data')

computing UMAP
    finished (0:00:27)
In [18]:
# Plot heatmap
sc.tl.dendrogram(adata, groupby='leiden', n_pcs = n_pcs, cor_method='pearson', linkage_method='complete', optimal_ordering=True)
sns.set(rc={'axes.facecolor':'white', 'axes.grid': False, 'axes.axisbelow': False, "xtick.labelbottom": True, 'ytick.labelleft': True, 'axes.linewidth': 1.0})
sc.pl.heatmap(adata, adata.var_names, groupby='leiden', swap_axes=True, cmap=seurat_style, dendrogram=True, vmin=-3.5, vmax=3.5,  show_gene_labels=True, figsize=(16, 10))

    using 'X_pca' with n_pcs = 20
Storing dendrogram info using `.uns['dendrogram_leiden']`

C:\Users\Xinrui\anaconda3\lib\site-packages\scanpy\plotting\_anndata.py:2414: FutureWarning: iteritems is deprecated and will be removed in a future version. Use .items instead.
  obs_tidy.index.value_counts(sort=False).iteritems()
In [19]:
adata.write(adata_fn, compression='gzip')

Rename clusters¶

Loading the clustering results¶

In [20]:
adata_fn = os.path.join(fig5path, 'Fig5_mouseBrain_10xFISHnCHIPs.h5ad')

adata = anndata.read(adata_fn)
display(adata)
display(adata.obs)
display(pd.DataFrame(adata.X, index=adata.obs_names, columns=adata.var_names))

leiden_ids = np.unique(adata.obs['leiden'])
print(leiden_ids)

AnnData object with n_obs × n_vars = 54834 × 53
    obs: 'x', 'y', 'leiden', 'QC', 'n_genes', 'normalize_total'
    var: 'n_cells', 'mean', 'std'
    uns: 'dendrogram_leiden', 'leiden', 'leiden_colors', 'log1p', 'neighbors', 'pca', 'umap'
    obsm: 'X_pca', 'X_umap'
    varm: 'PCs'
    obsp: 'connectivities', 'distances'
x y leiden QC n_genes normalize_total
1 7887.358491 13559.679250 10 cell 53 4985.189453
2 6411.150000 12702.433330 1 cell 53 2633.949951
3 6473.422414 12705.620690 1 cell 53 3883.879150
4 6685.226667 12703.000000 4 cell 53 6748.719727
5 6696.637500 12704.762500 0 cell 53 8492.500000
... ... ... ... ... ... ...
56347 6250.522989 1264.034483 6 cell 53 19507.751953
56348 5606.659259 1264.325926 13 cell 53 6758.881348
56349 5464.730769 1266.589744 7 cell 53 4381.282227
56350 5897.812500 1268.093750 13 cell 53 9357.312500
56351 2386.581081 354.770270 8 cell 53 2191.378174

54834 rows × 6 columns

M40 M37 M15 M47 M53 M21 M09 M34 M22 M49 ... M35 M31 M30 M24 M38 M39 M10 M26 M16 M08
1 -0.685691 -3.040047 0.015713 -0.360752 -1.929398 -0.892224 -1.407184 0.596070 1.049684 -0.423390 ... 2.290736 -0.101613 -1.074441 0.453871 -0.266024 -0.995892 -0.501919 1.638302 -0.404232 -1.252461
2 1.100721 1.557079 1.504797 -0.942026 -2.039713 1.600714 -0.345733 0.578655 0.160566 -1.078502 ... -0.064162 0.463169 0.248176 -0.205916 0.200232 -0.771157 -0.165907 -0.041364 0.163777 -1.816001
3 0.922584 1.624441 1.198860 -0.512981 -0.328418 1.523858 -0.678153 0.112750 -0.422142 0.202291 ... -0.167727 -0.046713 -0.201445 -0.210896 -0.245879 -0.443682 -0.231542 -0.105561 -0.369810 -0.564285
4 1.062692 0.314687 -0.801775 1.934706 1.401032 1.941626 1.032206 1.065207 1.712502 0.354619 ... 0.691696 0.165970 -0.032817 -0.703545 -0.297749 0.142018 -0.477113 -0.243606 -0.937429 -0.450291
5 0.318463 0.307173 -0.459689 0.631815 -0.127740 0.045705 -0.939245 -0.470986 0.494719 -0.042172 ... 1.227338 1.054807 0.507613 -0.084020 -0.545144 -0.174205 -0.842030 -0.152641 -1.126340 -0.894513
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
56347 -1.151868 0.015578 -0.775146 -0.463318 -0.348338 -0.169066 0.338190 -0.484735 -0.396904 -0.357825 ... 0.234742 0.921181 1.021010 -0.398054 -0.304159 -0.702721 -0.467647 -0.055781 -0.632381 -0.475379
56348 0.713704 0.449172 -0.011816 0.621240 0.468091 0.167068 0.614709 1.308247 0.959441 -0.225949 ... -0.750311 -1.094440 -0.606360 -0.520725 0.279993 0.109371 0.817338 -0.535278 0.117939 0.086753
56349 0.480499 -0.498654 0.413829 0.028349 0.754278 0.439535 1.030351 2.083175 0.855252 1.844793 ... -0.539726 0.029603 -0.014284 1.139597 1.197121 0.519468 1.271731 -0.519149 0.218924 0.672141
56350 -0.035127 -0.598269 -0.408276 1.792874 1.631192 0.056398 1.684673 2.125953 1.705631 -0.672530 ... -1.237041 -1.281118 -0.739867 -0.546828 -0.705438 -0.104619 0.255669 -0.504778 -0.500855 -1.226649
56351 0.540623 -1.548861 1.401411 0.362467 0.615148 0.193337 0.808659 1.764744 0.674784 1.878403 ... 2.073126 1.058152 -0.194620 0.570970 -0.175386 -0.015787 1.534887 0.783149 1.234711 2.086316

54834 rows × 53 columns

['0' '1' '10' '11' '12' '13' '14' '15' '16' '17' '2' '3' '4' '5' '6' '7'
 '8' '9']
In [21]:
cbm_scaled = pd.DataFrame(adata.X, index=adata.obs_names, columns=adata.var_names)
cbm_scaled.to_csv(os.path.join(fig5path, 'Figure5_FISHnCHIPs0316_scaledMatrix.csv'))
In [22]:
# Plot heatmap
sns.set(rc={'axes.facecolor':'white', 'axes.grid': False, 'axes.axisbelow': False, "xtick.labelbottom": True, 'ytick.labelleft': True, 'axes.linewidth': 1.0})
sc.pl.heatmap(adata, adata.var_names, groupby='leiden', swap_axes=True, cmap=seurat_style, dendrogram=True, vmin=-3.5, vmax=3.5,  show_gene_labels=True, figsize=(16, 10))

C:\Users\Xinrui\anaconda3\lib\site-packages\scanpy\plotting\_anndata.py:2414: FutureWarning: iteritems is deprecated and will be removed in a future version. Use .items instead.
  obs_tidy.index.value_counts(sort=False).iteritems()
In [23]:
# Assign names
class_names = {
    '0': 'Inhibitory neurons 1',
    '1': 'Astrocytes',
    '2': 'Mature oligodendrocytes 1',
    '3': 'Excitatory neurons 1',
    '4': 'Endothelial cells',
    '5': 'Excitatory neurons 2',
    '6': 'Excitatory neurons 3',
    '7': 'Blood vessel associated cells',
    '8': 'Unknown',
    '9':  'Excitatory neurons 4',
    '10': 'Mature oligodendrocytes 2',
    '11': 'Microglia',
    '12': 'Immature oligodendrocytes',
    '13': 'Pericytes',
    '14': 'Ependymal cells',
    '15': 'Excitatory neurons 5',
    '16': 'Excitatory neurons 6',
    '17': 'Inhibitory neurons 2'
}

adata.obs['class'] = adata.obs['leiden'].map(class_names)
In [24]:
display(adata.obs)
x y leiden QC n_genes normalize_total class
1 7887.358491 13559.679250 10 cell 53 4985.189453 Mature oligodendrocytes 2
2 6411.150000 12702.433330 1 cell 53 2633.949951 Astrocytes
3 6473.422414 12705.620690 1 cell 53 3883.879150 Astrocytes
4 6685.226667 12703.000000 4 cell 53 6748.719727 Endothelial cells
5 6696.637500 12704.762500 0 cell 53 8492.500000 Inhibitory neurons 1
... ... ... ... ... ... ... ...
56347 6250.522989 1264.034483 6 cell 53 19507.751953 Excitatory neurons 3
56348 5606.659259 1264.325926 13 cell 53 6758.881348 Pericytes
56349 5464.730769 1266.589744 7 cell 53 4381.282227 Blood vessel associated cells
56350 5897.812500 1268.093750 13 cell 53 9357.312500 Pericytes
56351 2386.581081 354.770270 8 cell 53 2191.378174 Unknown

54834 rows × 7 columns

In [25]:
adata.obs.to_csv(os.path.join(fig5path, 'cells_annotation.csv'))
In [26]:
# Plot clusters
layer = 'class'
n_cluster=len(np.unique(adata.obs[layer]))
print('%i clusters found!\n' % n_cluster)

cluster_order = np.unique(adata.obs.loc[:,layer])

# color_map_cluster = {cluster_order[i]: plot_color[i] for i in range(len(cluster_order))}
# print(color_map_cluster)
color_map_cluster = {
    'Astrocytes': '#F56867',
    'Blood vessel associated cells': '#ff478b',
    'Excitatory neurons 1': '#0095FF',
    'Excitatory neurons 2': '#ff9408',
    'Excitatory neurons 3': '#be03fd',
    'Excitatory neurons 4': '#009E73',
    'Excitatory neurons 5': '#ffff14',
    'Excitatory neurons 6': '#ee82ee',
    'Inhibitory neurons 2': '#aaff33',
    'Inhibitory neurons 1': '#b0c5ee',
    'Microglia': '#59BE86',
    'Pericytes': '#FF00AA',
    'Mature oligodendrocytes 1': '#ff000d',
    'Mature oligodendrocytes 2': '#1b6968',
    'Immature oligodendrocytes': '#fec848',
    'Ependymal cells':  '#21fc0d',
    'Endothelial cells': '#47ebe2' ,
    'Unknown': '#f7d1cf',
}

18 clusters found!
In [27]:
# Spatial
sns.set(rc={'axes.facecolor':'black', 'axes.grid': False, 'axes.axisbelow': False, 
            "xtick.labelbottom": False, 'ytick.labelleft': False, 
            'axes.linewidth': 1.0, 'axes.edgecolor': '.8'})

facet = sns.relplot(data=adata.obs, x="x", y="y", 
                    palette=color_map_cluster, 
                    col=layer, hue=layer, 
                    linewidth=0, col_wrap = 8, kind="scatter", 
                    s=3, aspect=spatial_aspect)
facet.fig.axes[0].invert_yaxis()
plt.savefig(clusters_idv_names_fn, bbox_inches='tight')

# Fused spatial
plt.figure(figsize=(12, 12))
sns.scatterplot(data = adata.obs, x = "x" , y = "y", s = 3, linewidth=0, hue= layer, palette=color_map_cluster)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0, ncol=1, facecolor='white')
plt.gca().set_aspect("equal")
plt.gca().invert_yaxis()
plt.savefig(clusters_fused_names_fn, bbox_inches='tight')
In [28]:
# Plot UMAP
sns.set(rc={'axes.facecolor':'white', 'axes.grid': False, 'figure.figsize': (10,10)})
sc.pl.umap(adata, color=layer, s=8, palette=color_map_cluster)
In [29]:
# Replot heatmap
from pandas.api.types import CategoricalDtype
from matplotlib.patches import Patch

cbm_plot = pd.DataFrame(adata.X, index=adata.obs_names, columns=adata.var_names)
cells_meta = adata.obs
In [30]:
layer = 'class'
cluster_order = np.array(['Immature oligodendrocytes',  'Mature oligodendrocytes 1', 'Mature oligodendrocytes 2', 
                          'Excitatory neurons 2','Excitatory neurons 4', 'Excitatory neurons 1', 'Excitatory neurons 3',  
                          'Excitatory neurons 5', 'Excitatory neurons 6', 'Inhibitory neurons 1',  'Inhibitory neurons 2',
                          'Astrocytes',  'Microglia',  'Pericytes',  'Blood vessel associated cells', 
                          'Endothelial cells',  'Ependymal cells', 'Unknown'])
cat_cluster = CategoricalDtype(cluster_order, ordered=True)
cells_meta.loc[:,layer] = cells_meta.loc[:,layer].astype(cat_cluster)
cells_meta.index = cells_meta.index.astype(int)
cells_meta = cells_meta.sort_index()
cells_meta = cells_meta.sort_values(layer)
print(cells_meta.shape)

cbm_plot.index = cbm_plot.index.astype(int)
cbm_plot = cbm_plot.loc[cells_meta.index.to_list(),:]
print(cbm_plot.shape)

row_cb = cells_meta.loc[:,layer].map(color_map_cluster) 
g = sns.clustermap(cbm_plot.T, col_cluster=False, row_cluster=True, 
                   col_colors=row_cb, cmap=seurat_style, vmin=-4.5, vmax=4.5,  standard_scale=None, 
                   xticklabels=False, yticklabels=True, figsize=(15, 10))
g.ax_heatmap.tick_params(axis='y', which='major', labelsize=8, labelrotation=0)
g.fig.subplots_adjust(bottom=0.01, top=0.98, right=0.8, left=0.01)
g.ax_cbar.set_position((0.05, .8, .03, .15))

handles = [Patch(facecolor=color_map_cluster[name]) for name in color_map_cluster]
plt.legend(handles, color_map_cluster, title=layer, bbox_to_anchor=(1.02, 0.8), bbox_transform=plt.gcf().transFigure, loc='best', fontsize=10)
plt.savefig(os.path.join(fig5path, 'Heatmap_class_4o5.png'), bbox_inches='tight')

C:\Users\Xinrui\AppData\Local\Temp\ipykernel_26456\1091199945.py:8: DeprecationWarning: In a future version, `df.iloc[:, i] = newvals` will attempt to set the values inplace instead of always setting a new array. To retain the old behavior, use either `df[df.columns[i]] = newvals` or, if columns are non-unique, `df.isetitem(i, newvals)`
  cells_meta.loc[:,layer] = cells_meta.loc[:,layer].astype(cat_cluster)

(54834, 7)
(54834, 53)
In [ ]: